Strong signal and real results. Worth committing a pilot to.
Cursor
The fastest-growing AI-native IDE with strong agent capabilities — evaluate it seriously, but watch for usage-based pricing surprises and cloud dependency.
DevTool·Agentic
cursor.comOur Take
What It Is
Cursor is an AI-native code editor built on VS Code that integrates AI agents directly into the development workflow. Multiple agents can run in parallel on the same project, with long-running agents persisting across sessions (v2.5+). The 2.6 release (March 2026) introduced MCP Apps — interactive UIs from Amplitude, Figma, and tldraw that render directly in chat — plus team marketplaces for private plugins.
Why It Matters
The numbers are hard to ignore: $2B ARR by March 2026, doubled in three months from $1B in November 2025. No developer tool has grown this fast. Cursor represents a bet that the IDE itself becomes the AI agent interface, not a separate terminal or chat window. The expansion to JetBrains IDEs via the Agent Client Protocol and cloud automations triggered by Slack, GitHub, and PagerDuty signals ambition beyond interactive coding.
For practitioners, the question is whether the agent-centric workflow genuinely improves productivity or just adds another layer of complexity. The 30+ marketplace plugins (Atlassian, Datadog, GitLab, Hugging Face) suggest an ecosystem forming, but the pricing shift to usage-based credits in June 2025 caught many developers off guard.
Key Developments
- Mar 2026: Cursor 2.6 with MCP Apps, team marketplaces, and Debug mode improvements.
- Feb 2026: Cursor 2.5 with long-running agents and Composer 1.5.
- Feb 2026: JetBrains IDE integration via Agent Client Protocol (ACP).
- Late 2025: Cloud automations triggered by external events (Slack, Linear, GitHub, PagerDuty).
- Mar 2026: TechCrunch reports $2B ARR milestone.
What to Watch
The multi-IDE expansion via ACP is the signal to track. If Cursor becomes a platform rather than a single editor, it changes the competitive landscape. Also watch for stabilisation of the usage-based pricing model — developers spending $44/month during coding sprints is creating friction that competitors like Claude Code (flat API pricing) and Copilot (subscription) can exploit.
Strengths
- Explosive market traction: $2B ARR by March 2026, 1M+ users, 360K paying customers. No developer tool has grown this fast.
- Agent-centric architecture: Multiple agents run in parallel on the same project with persistence across sessions.
- MCP Apps ecosystem: Interactive UIs from Figma, Amplitude, tldraw render directly in chat. 30+ partner plugins.
- Multi-IDE expansion: Available in IntelliJ, PyCharm, WebStorm via ACP, not just VS Code anymore.
Considerations
- Pricing unpredictability: Switched to $20 usage-based credits, effectively halving requests. Actual bills reach $44/month during intensive sprints.
- Code quality requires vigilance: AI sometimes introduces bugs, misses edge cases, or touches unintended files. Every change still needs human review.
- Cloud dependency: ~40% of developers encounter connectivity interruptions. Offline use reduces AI features by up to 80%.
- Privacy concerns: Code processed on external servers. Enterprise teams need to evaluate data handling policies carefully.
Resources
Articles
More in Developer Experience
Cursor· Gemini CLI· LiteLLM· Coding Agents· Google Antigravity· OpenRouter· Windsurf· Xcode Agentic Coding· Claude Code· GitHub Copilot· OpenAI Codex· Prompt Caching
Back to AI Radar